This is the conclusion from an interesting and thoughtful review by Morin and Grezes (2008, Clinical Neurophysiology, 38: 189-195). The basic observation is that the response of BA 44 (the posterior portion of Broca's area) does not discriminate between goal-directed and non-goal-directed actions, whereas BA 6 (a more posterior pre-motor site) does, responding more reliably across studies to goal-directed action. As macaque mirror neurons only respond goal-directed actions, this makes BA 6 the more likely candidate for the human homologue of monkey F5.
Of course the recent discovery of mirror neurons in monkey primary motor cortex (Tkach D, et al. 2007, Congruent activity during action and action observation in motor cortex. J Neurosci 27:13241-13250.) , kind of throws a monkey wrench into the whole mirror neuron enterprise. In the original reports did not find mirror neurons in M1 which was rightfully argued to be evidence against the possibility that the monkeys were covertly generating movement responses during the perception of actions. In other words, it ruled out the possibility that “mirror” responses were merely some kind of unimplemented motor reflex. This important "control" opened the door to a more interesting higher-level, function for mirror neurons. Now with the demonstration of “mirror” responses in low-level motor circuitry, it is entirely possible that “mirror” responses are nothing more than reflexive sensory-motor associations.